Effect of portable NIR instrument aging on fruit TSS prediction

There is a trend to take instrumentation from the laboratory to the field, e.g. spectrophotometers are commercially available for in field assessment of attributes of fruit on tree. Unfortunately, field users tend to place less emphasis on instrument maintenance, so understanding performance issues is important. Deterioration of lamp output quality over time and degradation of detector signal to noise ratio are issues associated with aging of an instrument. To document the effect of instrument aging on SWNIRS based fruit quality prediction, an experiment was conducted with several handheld PDA based spectrophotometers, with repeated spectra of a reference PFTE (Teflon) tile, and spectra of 20 apple fruit acquired at yearly intervals. Fruit TSS was also assessed, and used in development of a PLS regression models. The repeatability of each instrument was assessed as the standard deviation of absorbance of repeated measures of a reference, typically around 0.2 mAbs. Instrument changes were identified in performance and in PCA plots, but performance (apple TSS model) was not related to instrument repeatability. A piece-wise direct standardisation method is recommended to maintain multivariate calibration model across spectrometers.